I've built a fusor[1], and in its simplified form, it's just a large, negatively charged grid in a vacuum chamber with some hydrogen sitting around. Nothing to be afraid of!
You're absolutely right that these devices don't produce net energy. I built mine because I wanted to learn more about all the tech involved (high vacuum, high voltage, some nuclear physics). They're really good teaching projects for folks going into experimental physics, since there's a neat outcome without too much danger.
Each fusion event makes one He atom and releases about 17 MeV of fusion energy. Since (one mole) X (one eV) = ~100,000 Joules, fusing one mole (2 grams) of He would produce about 1.7 X 10^12 Joules. That's about 472 megawatt-hours of fusion energy produced.
Another commenter says that the energy INput (to the Farnsworth fusor) is about 100,000 times the energy OUTput; so fusing that 2 grams of He would require a 2,000 megawatt power plant to run for 23,600 hours, or 2.7 years.
No, deuterium-deuterium fusion reactions usually produce helium-3 and a neutron. Due to the conservation of energy, producing helium-4 requires the emission of a gamma ray. This happens rarely because, since the strong nuclear force is stronger than the electromagnetic force at small distances, fusion reactions tend to release energy as protons and neutrons rather than gamma rays.
You're absolutely right that these devices don't produce net energy. I built mine because I wanted to learn more about all the tech involved (high vacuum, high voltage, some nuclear physics). They're really good teaching projects for folks going into experimental physics, since there's a neat outcome without too much danger.
I assembled a coffee table book with some pretty pictures about it, PDF copy here: http://shia.wsyntax.com/stuff/fusion/fusor_book_web.pdf
[1] http://fusion.wsyntax.com/ , http://fusion.wsyntax.com/setup/